{
  "markdown": "# Desktop Control Skill\n\n🤖 **AI Agent Skill** for desktop automation using PyAutoGUI.\n\nControl mouse, keyboard, and screen programmatically through a simple CLI interface.\n\n## ⚡ Quick Install\n\n### For AI Agents\n\nAdd this skill to your AI coding agent with a single command:\n\n```bash\nnpx skills add patrickporto/desktop-agent\n```\n\n### For Users\n\nInstall the CLI with pipx (recommended):\n\n```bash\npipx install desktop-agent\n```\n\nOr run without installing using uvx:\n\n```bash\nuvx desktop-agent --help\n```\n\nOr using pip:\n\n```bash\npip install desktop-agent\n```\n\n---\n\n## 🎯 For AI Agents\n\nThis project is packaged as an **AI Agent Skill**. To use it:\n\n1. **Read the skill documentation**: [SKILL.md](SKILL.md)\n2. **Install the CLI**: Use `pip install desktop-agent` or `pipx install desktop-agent`\n3. **Invoke commands**: Use `desktop-agent <category> <command>`\n\n**Quick Reference for Agents:**\n- All commands documented in [SKILL.md](SKILL.md)\n- Practical examples in [examples/automation_examples.md](examples/automation_examples.md)\n- Help system: `desktop-agent --help`\n- All commands return JSON output by default\n\n---\n\n## 📦 Installation\n\n### Using pipx (Recommended)\n\n```bash\npipx install desktop-agent\n```\n\n### Using pip\n\n```bash\npip install desktop-agent\n```\n\n### Run without installing (using uvx)\n\n```bash\nuvx desktop-agent\n```\n\n\n## Usage\n\nThe CLI is organized into command categories:\n\n### 🖱️ Mouse (`mouse`)\n\n```bash\n# Move mouse to coordinates\ndesktop-agent mouse move 100 200\n\n# Move with duration (animation)\ndesktop-agent mouse move 100 200 --duration 1.0\n\n# Click at current position\ndesktop-agent mouse click\n\n# Click at specific coordinates\ndesktop-agent mouse click 500 500\n\n# Right click\ndesktop-agent mouse right-click\n\n# Double click\ndesktop-agent mouse double-click 300 400\n\n# Drag to coordinates\ndesktop-agent mouse drag 200 300\n\n# Scroll (positive = up, negative = down)\ndesktop-agent mouse scroll 5\ndesktop-agent mouse scroll -3\n\n# Get current mouse position\ndesktop-agent mouse position\n```\n\n### ⌨️ Keyboard (`keyboard`)\n\n```bash\n# Write text\ndesktop-agent keyboard write \"Hello World\"\n\n# Write with interval between keys\ndesktop-agent keyboard write \"Slow typing\" --interval 0.1\n\n# Press a key\ndesktop-agent keyboard press enter\n\n# Press multiple times\ndesktop-agent keyboard press a --presses 5\n\n# Execute keyboard shortcut\ndesktop-agent keyboard hotkey \"ctrl,c\"\ndesktop-agent keyboard hotkey \"ctrl,shift,esc\"\n\n# Hold/release key\ndesktop-agent keyboard keydown shift\ndesktop-agent keyboard keyup shift\n```\n\n### 🖼️ Screen (`screen`)\n\n```bash\n# Capture screenshot (full screen)\ndesktop-agent screen screenshot my_screen.png\n\n# Take screenshot of active window\ndesktop-agent screen screenshot active_window.png --active\n\n# Take screenshot of specific window\ndesktop-agent screen screenshot notepad.png --window \"Notepad\"\n\n# Screenshot of specific region (x,y,width,height)\ndesktop-agent screen screenshot region.png --region \"100,100,500,400\"\n\n# Locate image within active window\ndesktop-agent screen locate button.png --active\n\n# Locate center of image on screen\ndesktop-agent screen locate-center button.png --confidence 0.8\n\n# Find text coordinates within active window\ndesktop-agent screen locate-text-coordinates \"OK\" --active\n\n# Find text in specific image\ndesktop-agent screen locate-text-coordinates \"Confirm\" --image screenshot.png\n\n# Case-sensitive search\ndesktop-agent screen locate-text-coordinates \"Login\" --case-sensitive\n\n# Read all text from screen\ndesktop-agent screen read-all-text\n\n# Read text from image\ndesktop-agent screen read-all-text --image capture.png\n\n# Specify languages for OCR (default: pt,en)\ndesktop-agent screen locate-text-coordinates \"Button\" --lang \"en\"\n```\n\n### 💬 Messages (`message`)\n\n```bash\n# Show alert\ndesktop-agent message alert \"Hello!\"\n\n# Confirmation\ndesktop-agent message confirm \"Are you sure?\"\n\n# Input prompt\ndesktop-agent message prompt \"Enter your name:\"\n\n# Password\ndesktop-agent message password \"Enter your password:\"\n```\n\n### 📱 Applications (`app`)\n\n```bash\n# Open an application (cross-platform)\ndesktop-agent app open notepad\ndesktop-agent app open \"Google Chrome\"\n\n# Open with arguments\ndesktop-agent app open chrome --arg \"https://google.com\"\n\n# Focus on a window by title\ndesktop-agent app focus \"Untitled - Notepad\"\n\n# List all visible windows\ndesktop-agent app list\n```\n\n## Automation Examples\n\n### Open Notepad and write\n\n```bash\ndesktop-agent app open notepad\ndesktop-agent app focus notepad\ndesktop-agent keyboard write \"Hello from Desktop Skill!\"\n```\n\n### Capture screenshot and analyze\n\n```bash\ndesktop-agent screen screenshot full_screen.png\ndesktop-agent screen pixel 500 500\n```\n\n## Available Commands\n\nRun `desktop-agent --help` to see all commands:\n\n```bash\ndesktop-agent --help\ndesktop-agent mouse --help\ndesktop-agent keyboard --help\ndesktop-agent screen --help\ndesktop-agent message --help\n```\n\n## Project Structure\n\n```\ndesktop-skill/\n├── desktop_agent/       # Main package\n│   ├── __init__.py\n│   ├── commands/        # Command modules\n│   │   ├── __init__.py\n│   │   ├── mouse.py    # Mouse commands\n│   │   ├── keyboard.py # Keyboard commands\n│   │   ├── screen.py   # Screen/screenshot/OCR commands\n│   │   └── message.py  # Message boxes\n├── pyproject.toml      # Project configuration\n└── README.md           # This documentation\n```\n\n## Technologies\n\n- **PyAutoGUI**: GUI automation\n- **EasyOCR**: Optical character recognition\n- **Typer**: Modern CLI framework\n",
  "bytes": 5476,
  "sha": "ae4e9a5cc4916d6b5efb506804f5f3461fdeead60ab0f820c8369607f9e0a73f",
  "repo_slug": "patrickporto/desktop-agent",
  "fonte": "repo",
  "truncated": false,
  "api": "https://agentalog.com/api/listings/skl_patrickporto_desktop_agent_desktop_contr_ff3a4f6e/readme"
}